注意：本文源码是JDK8的版本，与之前的版本有较大差异
一、简介
ConcurrentHashMap是HashMap的线程安全实现版本，关于HashMap的源码分析，我们在《HashMap深入剖析（JDK8）》一文中有深入的剖析；ConcurrentHashMap由于其线程安全以及相对于Hashtable的锁更轻量级等相关特性，ConcurrentHashMap被广泛用于Dubbo，Spring等大部分Java开源框架之中；
我们先来看下ConcurrentHashMap类的注释：
/**
* A hash table supporting full concurrency of retrievals and
* high expected concurrency for updates(支持高并发的检索和更新). This class obeys the
* same functional specification as {@link java.util.Hashtable}, and
* includes versions of methods corresponding to each method of
* {@code Hashtable}. However, even though all operations are
* thread-safe, retrieval operations do <em>not</em> entail locking(尽管所有的操作都是线程安全的，检索操作也不需要加锁),
* and there is <em>not</em> any support for locking the entire table
* in a way that prevents all access(不需要锁住所有的记录数来阻止其他任何操作). This class is fully
* interoperable with {@code Hashtable} in programs that rely on its
* thread safety but not on its synchronization details.
*
* <p>Retrieval operations (including {@code get}) generally do not
* block(get操作时非阻塞的), so may overlap with update operations (including {@code put}
* and {@code remove}). Retrievals reflect the results of the most
* recently <em>completed</em> update operations holding upon their
* onset(检索出来的结果是最近更新的值). (More formally, an update operation for a given key bears a
* <em>happens-before</em> relation with any (non-null) retrieval for
* that key reporting the updated value.) For aggregate operations
* such as {@code putAll} and {@code clear}, concurrent retrievals may
* reflect insertion or removal of only some entries. Similarly,
* Iterators, Spliterators and Enumerations return elements reflecting the
* state of the hash table at some point at or since the creation of the
* iterator/enumeration. They do <em>not</em> throw {@link
* java.util.ConcurrentModificationException ConcurrentModificationException}.
* However, iterators are designed to be used by only one thread at a time.
* Bear in mind that the results of aggregate status methods including
* {@code size}, {@code isEmpty}, and {@code containsValue} are typically
* useful only when a map is not undergoing concurrent updates in other threads.
* Otherwise the results of these methods reflect transient states
* that may be adequate for monitoring or estimation purposes, but not
* for program control.(记住一些关于统计的方法，最好是在单线程的环境下使用，不然它只满足监控和估算的目的，在项目中使用它是无法精确返回的)
*
* <p>The table is dynamically expanded when there are too many
* collisions (当有太多的散列碰撞时，该表会动态增长)(i.e., keys that have distinct hash codes but fall into
* the same slot modulo the table size), with the expected average
* effect of maintaining roughly two bins per mapping (corresponding
* to a 0.75 load factor threshold for resizing). There may be much
* variance around this average as mappings are added and removed, but
* overall, this maintains a commonly accepted time/space tradeoff for
* hash tables. However, resizing this or any other kind of hash
* table may be a relatively slow operation. When possible, it is a
* good idea to provide a size estimate as an optional {@code
* initialCapacity} constructor argument. An additional optional
* {@code loadFactor} constructor argument provides a further means of
* customizing initial table capacity by specifying the table density
* to be used in calculating the amount of space to allocate for the
* given number of elements(再散列（扩容）是一件非常耗费资源的操作，最好是提前计算放入容器中有多少的元素来手动初始化装载因子和初始容量，这样会好很多). Also, for compatibility with previous
* versions of this class, constructors may optionally specify an
* expected {@code concurrencyLevel} as an additional hint for
* internal sizing. Note that using many keys with exactly the same
* {@code hashCode()} is a sure way to slow down performance of any
* hash table. To ameliorate impact, when keys are {@link Comparable},
* this class may use comparison order among keys to help break ties(当有很多的key的hashcode相等时会非常影响性能的(散列冲突)，key实现Comparable接口(自定义比较key)会好一点).
*
* <p>A {@link Set} projection of a ConcurrentHashMap may be created
* (using {@link #newKeySet()} or {@link #newKeySet(int)}), or viewed
* (using {@link #keySet(Object)} when only keys are of interest, and the
* mapped values are (perhaps transiently) not used or all take the
* same mapping value.
*
* <p>A ConcurrentHashMap can be used as scalable frequency map (a
* form of histogram or multiset) by using {@link
* java.util.concurrent.atomic.LongAdder} values and initializing via (能被用来频繁改变的Map，通过LongAdder)
* {@link #computeIfAbsent computeIfAbsent}. For example, to add a count
* to a {@code ConcurrentHashMap<String,LongAdder> freqs}, you can use
* {@code freqs.computeIfAbsent(k -> new LongAdder()).increment();}
*
* <p>This class and its views and iterators implement all of the
* <em>optional</em> methods of the {@link Map} and {@link Iterator}
* interfaces.(实现了Map和Iterator的所有方法)
*
* <p>Like {@link Hashtable} but unlike {@link HashMap}, this class
* does <em>not</em> allow {@code null} to be used as a key or value.(ConcurrentHashMap不允许key或者value为null)
*
* <p>ConcurrentHashMaps support a set of sequential and parallel bulk (ConcurrentHashMap提供支持批量操作)
* operations that, unlike most {@link Stream} methods, are designed
* to be safely, and often sensibly, applied even with maps that are
* being concurrently updated by other threads; for example, when
* computing a snapshot summary of the values in a shared registry.
* There are three kinds of operation, each with four forms, accepting
* functions with Keys, Values, Entries, and (Key, Value) arguments
* and/or return values. Because the elements of a ConcurrentHashMap
* are not ordered in any particular way, and may be processed in
* different orders in different parallel executions, the correctness
* of supplied functions should not depend on any ordering, or on any
* other objects or values that may transiently change while
* computation is in progress; and except for forEach actions, should
* ideally be side-effect-free. Bulk operations on {@link java.util.Map.Entry}
* objects do not support method {@code setValue}.
*
* <ul>
* <li> forEach: Perform a given action on each element.
* A variant form applies a given transformation on each element
* before performing the action.</li>
*
* <li> search: Return the first available non-null result of
* applying a given function on each element; skipping further
* search when a result is found.</li>
*
* <li> reduce: Accumulate each element. The supplied reduction
* function cannot rely on ordering (more formally, it should be
* both associative and commutative). There are five variants:
*
* <ul>
*
* <li> Plain reductions. (There is not a form of this method for
* (key, value) function arguments since there is no corresponding
* return type.)</li>
*
* <li> Mapped reductions that accumulate the results of a given
* function applied to each element.</li>
*
* <li> Reductions to scalar doubles, longs, and ints, using a
* given basis value.</li>
*
* </ul>
* </li>
* </ul>
*
* <p>These bulk operations accept a {@code parallelismThreshold}
* argument. Methods proceed sequentially if the current map size is
* estimated to be less than the given threshold. Using a value of
* {@code Long.MAX_VALUE} suppresses all parallelism. Using a value
* of {@code 1} results in maximal parallelism by partitioning into
* enough subtasks to fully utilize the {@link
* ForkJoinPool#commonPool()} that is used for all parallel
* computations. Normally, you would initially choose one of these
* extreme values, and then measure performance of using in-between
* values that trade off overhead versus throughput.
*
* <p>The concurrency properties of bulk operations follow
* from those of ConcurrentHashMap: Any non-null result returned
* from {@code get(key)} and related access methods bears a
* happens-before relation with the associated insertion or
* update. The result of any bulk operation reflects the
* composition of these per-element relations (but is not
* necessarily atomic with respect to the map as a whole unless it
* is somehow known to be quiescent). Conversely, because keys
* and values in the map are never null, null serves as a reliable
* atomic indicator of the current lack of any result. To
* maintain this property, null serves as an implicit basis for
* all non-scalar reduction operations. For the double, long, and
* int versions, the basis should be one that, when combined with
* any other value, returns that other value (more formally, it
* should be the identity element for the reduction). Most common
* reductions have these properties; for example, computing a sum
* with basis 0 or a minimum with basis MAX_VALUE.
*
* <p>Search and transformation functions provided as arguments
* should similarly return null to indicate the lack of any result
* (in which case it is not used). In the case of mapped
* reductions, this also enables transformations to serve as
* filters, returning null (or, in the case of primitive
* specializations, the identity basis) if the element should not
* be combined. You can create compound transformations and
* filterings by composing them yourself under this "null means
* there is nothing there now" rule before using them in search or
* reduce operations.
*
* <p>Methods accepting and/or returning Entry arguments maintain
* key-value associations. They may be useful for example when
* finding the key for the greatest value. Note that "plain" Entry
* arguments can be supplied using {@code new
* AbstractMap.SimpleEntry(k,v)}.
*
* <p>Bulk operations may complete abruptly, throwing an
* exception encountered in the application of a supplied
* function. Bear in mind when handling such exceptions that other
* concurrently executing functions could also have thrown
* exceptions, or would have done so if the first exception had
* not occurred.
*
* <p>Speedups for parallel compared to sequential forms are common
* but not guaranteed. Parallel operations involving brief functions
* on small maps may execute more slowly than sequential forms if the
* underlying work to parallelize the computation is more expensive
* than the computation itself. Similarly, parallelization may not
* lead to much actual parallelism if all processors are busy
* performing unrelated tasks.
*
* <p>All arguments to all task methods must be non-null.
*
* <p>This class is a member of the
* <a href="{@docRoot}/../technotes/guides/collections/index.html">
* Java Collections Framework</a>.
*
* @since 1.5
* @author Doug Lea
* @param <K> the type of keys maintained by this map
* @param <V> the type of mapped values
*/
其中一部分已经简单的用中文翻译了一下，基于以上类文件的注释，我们能简单得出如下结论：
- ConcurrentHashMap支持高并发的访问和更新，它是线程安全的
- 检索操作不用加锁，get方法是非阻塞的
- key和value都不允许为null
- key可以实现自定义的Comparable接口
1.1 ConcurrentHashMap类继承关系
我们先从Map的族谱中来概览一下ConcurrentHashMap的继承关系，如图：
ConcurrentHashMap类继承关系
ConcurrentHashMap 继承了AbstractMap ,并且实现了ConcurrentMap接口。与HashMap比对：
- 相同点：都继承了AbstractMap接口
- 不同点：HashMap实现了Map接口，ConcurrentHashMap实现了ConcurrentMap接口，而ConcurrentMap继承了Map接口，使用default关键字定义了一些方法 。
从继承关系上看ConcurrentHashMap与HashMap并没有太大的区别。
ConcurrentMap接口提供的功能：
| 方法签名 | 功能 |
|---|---|
| getOrDefault(Object key, V defaultValue) | 返回指定key对应的值；如果Map不存在该key，则返回defaultValue |
| forEach(BiConsumer action) | 遍历Map的所有Entry，并对其进行指定的aciton操作 |
| putIfAbsent(K key, V value) | 如果Map不存在指定的key，则插入<K,V>；否则，直接返回该key对应的值 |
| remove(Object key, Object value) | 删除与<key,value>完全匹配的Entry，并返回true；否则，返回false |
| replace(K key, V oldValue, V newValue) | 如果存在key，且值和oldValue一致，则更新为newValue，并返回true；否则，返回false |
| replace(K key, V value) | 如果存在key，则更新为value，返回旧value；否则，返回null |
| replaceAll(BiFunction function) | 遍历Map的所有Entry，并对其进行指定的funtion操作 |
| computeIfAbsent(K key, Function mappingFunction) | 如果Map不存在指定的key，则通过mappingFunction计算value并插入 |
| computeIfPresent(K key, BiFunction remappingFunction) | 如果Map存在指定的key，则通过mappingFunction计算value并替换旧值 |
| compute(K key, BiFunction remappingFunction) | 根据指定的key，查找value；然后根据得到的value和remappingFunction重新计算新值，并替换旧值 |
| merge(K key, V value, BiFunction remappingFunction) | 如果key不存在，则插入value；否则，根据key对应的值和remappingFunction计算新值，并替换旧值 |
1.2 ConcurrentHashMap实现原理概述
从上文可以知道，ConcurrentHashMap是一个线程安全的Map集合，可以应对高并发的场景，保证线程安全。相比较HashTable，它的锁粒度更加的细化，因为HashTable的方法都是用Synchronized修饰的，效率非常的低下。JDK1.8之前ConcurrentHashMap使用锁分段技术，将数据分成一段段的存储，每一个数据段配置一把锁，相互之间不影响，而JDK1.8开始摒弃了Segment(锁段)的概念，启用了全新的实现，也就是利用CAS(会在并发相关的文章中讲述其原理)+Synchronized来保证并发更新的安全，底层采用的跟JDK1.8中HashMap类似的“数组+链表+红黑树”的存储结构，但是为了做到并发，又增加了很多辅助的类，例如TreeBin，Traverser等对象内部类。其结构如下：
ConcurrentHashMap结构(jdk1.8).png
基本结构：
ConcurrentHashMap内部维护了一个Node类型的数组，也就是table：
/**
* The array of bins. Lazily initialized upon first insertion.
* Size is always a power of two. Accessed directly by iterators.
*/
transient volatile Node<K,V>[] table;
数组的每一个位置table[i]代表了一个桶，当插入键值对时，会根据键的hash值映射到不同的桶位置，table一共可以包含4种不同类型的桶：Node、TreeBin、ForwardingNode、ReservationNode。上图中，不同的桶用不同颜色表示。可以看到，有的桶链接着链表，有的桶链接着树。
需要注意的是：TreeBin所链接的是一颗红黑树，红黑树的节点用TreeNode表示，所以ConcurrentHashMap中实际上一共有5种不同类型的Node节点。
之所以用TreeBin而不是直接用TreeNode，是因为红黑树的操作比较复杂，包括构建、左旋、右旋、删除、平衡等操作，用一个代理节点TreeBin来包含这些复杂操作，其实是一种“职责分离”的思想。另外TreeBin中也包含了一些加/解锁的操作。
二、源码剖析
2.1 ConcurrentHashMap内部类
ConcurrentHashMap包含了很多内部类，其中主要的内部类框架图如下图所示：
ConcurrentHashMap内部类之Node.png
ConcurrentHashMap内部类之BulkTask.png
从上面的两幅图中可以看到，ConcurrentHashMap的内部类非常的庞大，第二个图是在JDK1.8下增加的类，下面对其中主要的内部类进行分析和讲解。
1. Node类
Node类主要用于存储具体键值对，其子类有ForwardingNode、ReservationNode、TreeNode和TreeBin四个子类。四个子类具体的代码在之后的具体例子中进行分析讲解。
2. Traverser类
Traverser类主要用于遍历操作，其子类有BaseIterator、KeySpliterator、ValueSpliterator、EntrySpliterator四个类，BaseIterator用于遍历操作。KeySplitertor、ValueSpliterator、EntrySpliterator则用于键、值、键值对的划分。
3. CollectionView类
CollectionView抽象类主要定义了视图操作，其子类KeySetView、ValueSetView、EntrySetView分别表示键视图、值视图、键值对视图。对视图均可以进行操作。
2.2 节点定义
从上面我们可以知道ConcurrentHashMap中有5种节点，我们现在分别来看下具体的每种节点的定义和作用。
2.2.1 Node节点
Node节点的定义比较简单，主要用来存放元素的key,value,hash值,next下一个链表节点的引用数据，它也是其他四种类型节点的父类。链表具体原理：
默认链接到table[i]——桶上的节点就是Node节点。
当出现hash冲突时，Node节点会首先以链表的形式链接到table上，当节点数量超过一定数目时，链表会转化为红黑树。因为链表查找的平均时间复杂度为O(n)，而红黑树是一种平衡二叉树，其平均时间复杂度为O(logn)。
/**
* Key-value entry. This class is never exported out as a
* user-mutable Map.Entry (i.e., one supporting setValue; see
* MapEntry below), but can be used for read-only traversals used
* in bulk tasks. Subclasses of Node with a negative hash field
* are special, and contain null keys and values (but are never
* exported). Otherwise, keys and vals are never null. (普通的Entry节点, 以链表形式保存时才会使用, 存储实际的数据.)
*/
static class Node<K,V> implements Map.Entry<K,V> {
final int hash;
final K key;
volatile V val;
volatile Node<K,V> next; // 链表指针
Node(int hash, K key, V val, Node<K,V> next) {
this.hash = hash;
this.key = key;
this.val = val;
this.next = next;
}
public final K getKey() { return key; }
public final V getValue() { return val; }
public final int hashCode() { return key.hashCode() ^ val.hashCode(); }
public final String toString(){ return key + "=" + val; }
public final V setValue(V value) {
throw new UnsupportedOperationException();
}
public final boolean equals(Object o) {
Object k, v, u; Map.Entry<?,?> e;
return ((o instanceof Map.Entry) &&
(k = (e = (Map.Entry<?,?>)o).getKey()) != null &&
(v = e.getValue()) != null &&
(k == key || k.equals(key)) &&
(v == (u = val) || v.equals(u)));
}
/**
* Virtualized support for map.get(); overridden in subclasses. 链表查找
*/
Node<K,V> find(int h, Object k) {
Node<K,V> e = this;
if (k != null) {
do {
K ek;
if (e.hash == h &&
((ek = e.key) == k || (ek != null && k.equals(ek))))
return e;
} while ((e = e.next) != null);
}
return null;
}
}
2.2.2 TreeNode节点
TreeNode就是红黑树的节点，存放了父节点，左子节点，右子节点的引用，以及红黑节点标识，TreeNode不会直接链接到table[i]——桶上面，而是由TreeBin链接，TreeBin会指向红黑树的根节点。
/**
* Nodes for use in TreeBins 红黑树节点, 存储实际的数据.
*/
static final class TreeNode<K,V> extends Node<K,V> {
TreeNode<K,V> parent; // red-black tree links
TreeNode<K,V> left;
TreeNode<K,V> right;
TreeNode<K,V> prev; // needed to unlink next upon deletion (prev指针是为了方便删除.删除链表的非头节点时，需要知道它的前驱节点才能删除，所以直接提供一个prev指针)
boolean red;
TreeNode(int hash, K key, V val, Node<K,V> next,
TreeNode<K,V> parent) {
super(hash, key, val, next);
this.parent = parent;
}
Node<K,V> find(int h, Object k) {
return findTreeNode(h, k, null);
}
/**
* Returns the TreeNode (or null if not found) for the given key
* starting at given root. 以当前节点（this）为根节点，开始遍历查找指定key.
*/
final TreeNode<K,V> findTreeNode(int h, Object k, Class<?> kc) {
if (k != null) {
TreeNode<K,V> p = this;
do {
int ph, dir; K pk; TreeNode<K,V> q;
TreeNode<K,V> pl = p.left, pr = p.right;
if ((ph = p.hash) > h)
p = pl;
else if (ph < h)
p = pr;
else if ((pk = p.key) == k || (pk != null && k.equals(pk)))
return p;
else if (pl == null)
p = pr;
else if (pr == null)
p = pl;
else if ((kc != null ||
(kc = comparableClassFor(k)) != null) &&
(dir = compareComparables(kc, k, pk)) != 0)
p = (dir < 0) ? pl : pr;
else if ((q = pr.findTreeNode(h, k, kc)) != null)
return q;
else
p = pl;
} while (p != null);
}
return null;
}
}
2.2.3 TreeBin节点
TreeBin相当于TreeNode的代理节点。TreeBin会直接链接到table[i]----桶上面，该节点提供了一系列红黑树相关的操作，以及加锁、解锁操作。内部属性有root，first节点，以及root节点的锁状态变量lockState，这是一个读写锁的状态。用于存放红黑树的root节点，并用读写锁lockState控制在写操作即将要调整树结构前，先让读线程完成读操作。从链表结构调整为红黑树时，table中索引下标存储的即为TreeBin。
/** TreeNode的代理节点（相当于封装了TreeNode的容器，提供针对红黑树的转换操作和锁控制）hash值固定为-3
* TreeNodes used at the heads of bins. TreeBins do not hold user
* keys or values, but instead point to list of TreeNodes and
* their root. They also maintain a parasitic read-write lock
* forcing writers (who hold bin lock) to wait for readers (who do
* not) to complete before tree restructuring operations.
*/
static final class TreeBin<K,V> extends Node<K,V> {
TreeNode<K,V> root; // 红黑树结构的根节点
volatile TreeNode<K,V> first; // 链表结构的头节点
volatile Thread waiter; // 最近的一个设置WAITER标识位的线程
volatile int lockState; // 整体的锁状态标识位
// values for lockState
static final int WRITER = 1; // set while holding write lock 二进制001，红黑树的写锁状态
static final int WAITER = 2; // set when waiting for write lock 二进制010，红黑树的等待获取写锁状态
static final int READER = 4; // increment value for setting read lock 二进制100，红黑树的读锁状态，读可以并发，每多一个读线程，lockState都加上一个READER值
/** 在hashCode相等并且不是Comparable类型时，用此方法判断大小.
* Tie-breaking utility for ordering insertions when equal
* hashCodes and non-comparable. We don't require a total
* order, just a consistent insertion rule to maintain
* equivalence across rebalancings. Tie-breaking further than
* necessary simplifies testing a bit.
*/
static int tieBreakOrder(Object a, Object b) {
int d;
if (a == null || b == null ||
(d = a.getClass().getName().
compareTo(b.getClass().getName())) == 0)
d = (System.identityHashCode(a) <= System.identityHashCode(b) ?
-1 : 1);
return d;
}
/** 将以b为头节点的链表转换为红黑树.
* Creates bin with initial set of nodes headed by b.
*/
TreeBin(TreeNode<K,V> b) {
super(TREEBIN, null, null, null);
this.first = b;
TreeNode<K,V> r = null;
for (TreeNode<K,V> x = b, next; x != null; x = next) {
next = (TreeNode<K,V>)x.next;
x.left = x.right = null;
if (r == null) {
x.parent = null;
x.red = false;
r = x;
}
else {
K k = x.key;
int h = x.hash;
Class<?> kc = null;
for (TreeNode<K,V> p = r;;) {
int dir, ph;
K pk = p.key;
if ((ph = p.hash) > h)
dir = -1;
else if (ph < h)
dir = 1;
else if ((kc == null &&
(kc = comparableClassFor(k)) == null) ||
(dir = compareComparables(kc, k, pk)) == 0)
dir = tieBreakOrder(k, pk);
TreeNode<K,V> xp = p;
if ((p = (dir <= 0) ? p.left : p.right) == null) {
x.parent = xp;
if (dir <= 0)
xp.left = x;
else
xp.right = x;
r = balanceInsertion(r, x);
break;
}
}
}
}
this.root = r;
assert checkInvariants(root);
}
/** 对红黑树的根节点加写锁.
* Acquires write lock for tree restructuring.
*/
private final void lockRoot() {
if (!U.compareAndSwapInt(this, LOCKSTATE, 0, WRITER))
contendedLock(); // offload to separate method
}
/** 释放写锁.
* Releases write lock for tree restructuring.
*/
private final void unlockRoot() {
lockState = 0;
}
/**
* Possibly blocks awaiting root lock.
*/
private final void contendedLock() {
boolean waiting = false;
for (int s;;) {
if (((s = lockState) & ~WAITER) == 0) {
if (U.compareAndSwapInt(this, LOCKSTATE, s, WRITER)) {
if (waiting)
waiter = null;
return;
}
}
else if ((s & WAITER) == 0) {
if (U.compareAndSwapInt(this, LOCKSTATE, s, s | WAITER)) {
waiting = true;
waiter = Thread.currentThread();
}
}
else if (waiting)
LockSupport.park(this);
}
}
/** 从根节点开始遍历查找，找到“相等”的节点就返回它，没找到就返回null, 当存在写锁时，以链表方式进行查找
* Returns matching node or null if none. Tries to search
* using tree comparisons from root, but continues linear
* search when lock not available.
*/
final Node<K,V> find(int h, Object k) {
if (k != null) {
for (Node<K,V> e = first; e != null; ) {
int s; K ek;
/**
* 两种特殊情况下以链表的方式进行查找:
* 1. 有线程正持有写锁，这样做能够不阻塞读线程
* 2. 有线程等待获取写锁，不再继续加读锁，相当于“写优先”模式
*/
if (((s = lockState) & (WAITER|WRITER)) != 0) {
if (e.hash == h &&
((ek = e.key) == k || (ek != null && k.equals(ek))))
return e;
e = e.next;
}
else if (U.compareAndSwapInt(this, LOCKSTATE, s,
s + READER)) {
TreeNode<K,V> r, p;
try {
p = ((r = root) == null ? null :
r.findTreeNode(h, k, null));
} finally {
Thread w;
if (U.getAndAddInt(this, LOCKSTATE, -READER) ==
(READER|WAITER) && (w = waiter) != null)
LockSupport.unpark(w);
}
return p;
}
}
}
return null;
}
/** 查找指定key对应的节点,如果未找到，则插入.
* Finds or adds a node.
* @return null if added 插入成功返回null, 否则返回找到的节点
*/
final TreeNode<K,V> putTreeVal(int h, K k, V v) {
Class<?> kc = null;
boolean searched = false;
for (TreeNode<K,V> p = root;;) {
int dir, ph; K pk;
if (p == null) {
first = root = new TreeNode<K,V>(h, k, v, null, null);
break;
}
else if ((ph = p.hash) > h)
dir = -1;
else if (ph < h)
dir = 1;
else if ((pk = p.key) == k || (pk != null && k.equals(pk)))
return p;
else if ((kc == null &&
(kc = comparableClassFor(k)) == null) ||
(dir = compareComparables(kc, k, pk)) == 0) {
if (!searched) {
TreeNode<K,V> q, ch;
searched = true;
if (((ch = p.left) != null &&
(q = ch.findTreeNode(h, k, kc)) != null) ||
((ch = p.right) != null &&
(q = ch.findTreeNode(h, k, kc)) != null))
return q;
}
dir = tieBreakOrder(k, pk);
}
TreeNode<K,V> xp = p;
if ((p = (dir <= 0) ? p.left : p.right) == null) {
TreeNode<K,V> x, f = first;
first = x = new TreeNode<K,V>(h, k, v, f, xp);
if (f != null)
f.prev = x;
if (dir <= 0)
xp.left = x;
else
xp.right = x;
if (!xp.red)
x.red = true;
else {
lockRoot();
try {
root = balanceInsertion(root, x);
} finally {
unlockRoot();
}
}
break;
}
}
assert checkInvariants(root);
return null;
}
/** 删除红黑树的节点：
* 1. 红黑树规模太小时，返回true，然后进行 树 -> 链表 的转化;
* 2. 红黑树规模足够时，不用变换成链表，但删除节点时需要加写锁.
* Removes the given node, that must be present before this
* call. This is messier than typical red-black deletion code
* because we cannot swap the contents of an interior node
* with a leaf successor that is pinned by "next" pointers
* that are accessible independently of lock. So instead we
* swap the tree linkages.
*
* @return true if now too small, so should be untreeified
*/
final boolean removeTreeNode(TreeNode<K,V> p) {
TreeNode<K,V> next = (TreeNode<K,V>)p.next;
TreeNode<K,V> pred = p.prev; // unlink traversal pointers
TreeNode<K,V> r, rl;
if (pred == null)
first = next;
else
pred.next = next;
if (next != null)
next.prev = pred;
if (first == null) {
root = null;
return true;
}
if ((r = root) == null || r.right == null || // too small
(rl = r.left) == null || rl.left == null)
return true;
lockRoot();
try {
TreeNode<K,V> replacement;
TreeNode<K,V> pl = p.left;
TreeNode<K,V> pr = p.right;
if (pl != null && pr != null) {
TreeNode<K,V> s = pr, sl;
while ((sl = s.left) != null) // find successor
s = sl;
boolean c = s.red; s.red = p.red; p.red = c; // swap colors
TreeNode<K,V> sr = s.right;
TreeNode<K,V> pp = p.parent;
if (s == pr) { // p was s's direct parent
p.parent = s;
s.right = p;
}
else {
TreeNode<K,V> sp = s.parent;
if ((p.parent = sp) != null) {
if (s == sp.left)
sp.left = p;
else
sp.right = p;
}
if ((s.right = pr) != null)
pr.parent = s;
}
p.left = null;
if ((p.right = sr) != null)
sr.parent = p;
if ((s.left = pl) != null)
pl.parent = s;
if ((s.parent = pp) == null)
r = s;
else if (p == pp.left)
pp.left = s;
else
pp.right = s;
if (sr != null)
replacement = sr;
else
replacement = p;
}
else if (pl != null)
replacement = pl;
else if (pr != null)
replacement = pr;
else
replacement = p;
if (replacement != p) {
TreeNode<K,V> pp = replacement.parent = p.parent;
if (pp == null)
r = replacement;
else if (p == pp.left)
pp.left = replacement;
else
pp.right = replacement;
p.left = p.right = p.parent = null;
}
root = (p.red) ? r : balanceDeletion(r, replacement);
if (p == replacement) { // detach pointers
TreeNode<K,V> pp;
if ((pp = p.parent) != null) {
if (p == pp.left)
pp.left = null;
else if (p == pp.right)
pp.right = null;
p.parent = null;
}
}
} finally {
unlockRoot();
}
assert checkInvariants(root);
return false;
}
/* ------------------------------------------------------------ */
// Red-black tree methods, all adapted from CLR // 以下是红黑树的经典操作方法，改编自《算法导论》
static <K,V> TreeNode<K,V> rotateLeft(TreeNode<K,V> root,
TreeNode<K,V> p) {
TreeNode<K,V> r, pp, rl;
if (p != null && (r = p.right) != null) {
if ((rl = p.right = r.left) != null)
rl.parent = p;
if ((pp = r.parent = p.parent) == null)
(root = r).red = false;
else if (pp.left == p)
pp.left = r;
else
pp.right = r;
r.left = p;
p.parent = r;
}
return root;
}
static <K,V> TreeNode<K,V> rotateRight(TreeNode<K,V> root,
TreeNode<K,V> p) {
TreeNode<K,V> l, pp, lr;
if (p != null && (l = p.left) != null) {
if ((lr = p.left = l.right) != null)
lr.parent = p;
if ((pp = l.parent = p.parent) == null)
(root = l).red = false;
else if (pp.right == p)
pp.right = l;
else
pp.left = l;
l.right = p;
p.parent = l;
}
return root;
}
static <K,V> TreeNode<K,V> balanceInsertion(TreeNode<K,V> root,
TreeNode<K,V> x) {
x.red = true;
for (TreeNode<K,V> xp, xpp, xppl, xppr;;) {
if ((xp = x.parent) == null) {
x.red = false;
return x;
}
else if (!xp.red || (xpp = xp.parent) == null)
return root;
if (xp == (xppl = xpp.left)) {
if ((xppr = xpp.right) != null && xppr.red) {
xppr.red = false;
xp.red = false;
xpp.red = true;
x = xpp;
}
else {
if (x == xp.right) {
root = rotateLeft(root, x = xp);
xpp = (xp = x.parent) == null ? null : xp.parent;
}
if (xp != null) {
xp.red = false;
if (xpp != null) {
xpp.red = true;
root = rotateRight(root, xpp);
}
}
}
}
else {
if (xppl != null && xppl.red) {
xppl.red = false;
xp.red = false;
xpp.red = true;
x = xpp;
}
else {
if (x == xp.left) {
root = rotateRight(root, x = xp);
xpp = (xp = x.parent) == null ? null : xp.parent;
}
if (xp != null) {
xp.red = false;
if (xpp != null) {
xpp.red = true;
root = rotateLeft(root, xpp);
}
}
}
}
}
}
static <K,V> TreeNode<K,V> balanceDeletion(TreeNode<K,V> root,
TreeNode<K,V> x) {
for (TreeNode<K,V> xp, xpl, xpr;;) {
if (x == null || x == root)
return root;
else if ((xp = x.parent) == null) {
x.red = false;
return x;
}
else if (x.red) {
x.red = false;
return root;
}
else if ((xpl = xp.left) == x) {
if ((xpr = xp.right) != null && xpr.red) {
xpr.red = false;
xp.red = true;
root = rotateLeft(root, xp);
xpr = (xp = x.parent) == null ? null : xp.right;
}
if (xpr == null)
x = xp;
else {
TreeNode<K,V> sl = xpr.left, sr = xpr.right;
if ((sr == null || !sr.red) &&
(sl == null || !sl.red)) {
xpr.red = true;
x = xp;
}
else {
if (sr == null || !sr.red) {
if (sl != null)
sl.red = false;
xpr.red = true;
root = rotateRight(root, xpr);
xpr = (xp = x.parent) == null ?
null : xp.right;
}
if (xpr != null) {
xpr.red = (xp == null) ? false : xp.red;
if ((sr = xpr.right) != null)
sr.red = false;
}
if (xp != null) {
xp.red = false;
root = rotateLeft(root, xp);
}
x = root;
}
}
}
else { // symmetric
if (xpl != null && xpl.red) {
xpl.red = false;
xp.red = true;
root = rotateRight(root, xp);
xpl = (xp = x.parent) == null ? null : xp.left;
}
if (xpl == null)
x = xp;
else {
TreeNode<K,V> sl = xpl.left, sr = xpl.right;
if ((sl == null || !sl.red) &&
(sr == null || !sr.red)) {
xpl.red = true;
x = xp;
}
else {
if (sl == null || !sl.red) {
if (sr != null)
sr.red = false;
xpl.red = true;
root = rotateLeft(root, xpl);
xpl = (xp = x.parent) == null ?
null : xp.left;
}
if (xpl != null) {
xpl.red = (xp == null) ? false : xp.red;
if ((sl = xpl.left) != null)
sl.red = false;
}
if (xp != null) {
xp.red = false;
root = rotateRight(root, xp);
}
x = root;
}
}
}
}
}
/**
* Recursive invariant check 递归检查红黑树的正确性
*/
static <K,V> boolean checkInvariants(TreeNode<K,V> t) {
TreeNode<K,V> tp = t.parent, tl = t.left, tr = t.right,
tb = t.prev, tn = (TreeNode<K,V>)t.next;
if (tb != null && tb.next != t)
return false;
if (tn != null && tn.prev != t)
return false;
if (tp != null && t != tp.left && t != tp.right)
return false;
if (tl != null && (tl.parent != t || tl.hash > t.hash))
return false;
if (tr != null && (tr.parent != t || tr.hash < t.hash))
return false;
if (t.red && tl != null && tl.red && tr != null && tr.red)
return false;
if (tl != null && !checkInvariants(tl))
return false;
if (tr != null && !checkInvariants(tr))
return false;
return true;
}
private static final sun.misc.Unsafe U;
private static final long LOCKSTATE;
static {
try {
U = sun.misc.Unsafe.getUnsafe();
Class<?> k = TreeBin.class;
LOCKSTATE = U.objectFieldOffset
(k.getDeclaredField("lockState"));
} catch (Exception e) {
throw new Error(e);
}
}
}
2.2.4 ForwardingNode节点
这是一个真正的辅助类，该类仅仅只存活在ConcurrentHashMap扩容操作时。只是一个标志节点，并且指向nextTable，它提供find方法而已。该类也是继承Node节点，其hash为-1，key、value、next均为null。如下：
/**
* 并不是我们传统的包含key-value的节点，只是一个标志节点，并且指向nextTable，提供find方法而已。
* 生命周期：仅存活于扩容操作且bin不为null时，一定会出现在每个bin的首位。
* A node inserted at head of bins during transfer operations.
*/
static final class ForwardingNode<K,V> extends Node<K,V> {
final Node<K,V>[] nextTable;
ForwardingNode(Node<K,V>[] tab) {
super(MOVED, null, null, null); // 此节点hash=-1，key、value、next均为null
this.nextTable = tab;
}
Node<K,V> find(int h, Object k) {
// loop to avoid arbitrarily deep recursion on forwarding nodes
// 查nextTable节点，outer避免深度递归
outer: for (Node<K,V>[] tab = nextTable;;) {
Node<K,V> e; int n;
if (k == null || tab == null || (n = tab.length) == 0 ||
(e = tabAt(tab, (n - 1) & h)) == null)
return null;
for (;;) { // CAS算法多和死循环搭配！直到查到或null
int eh; K ek;
if ((eh = e.hash) == h &&
((ek = e.key) == k || (ek != null && k.equals(ek))))
return e;
if (eh < 0) {
if (e instanceof ForwardingNode) {
tab = ((ForwardingNode<K,V>)e).nextTable;
continue outer;
}
else
return e.find(h, k);
}
if ((e = e.next) == null)
return null;
}
}
}
}
2.2.5 ReservationNode节点
保留节点，ConcurrentHashMap中的一些特殊方法会专门用到该类节点。
/**
* 保留节点.
* hash值固定为-3， 不保存实际数据
* 只在computeIfAbsent和compute这两个函数式API中充当占位符加锁使用
* A place-holder node used in computeIfAbsent and compute
*/
static final class ReservationNode<K,V> extends Node<K,V> {
ReservationNode() {
super(RESERVED, null, null, null);
}
Node<K,V> find(int h, Object k) {
return null;
}
}
2.3 ConcurrentHashMap构造器
2.3.1 构造器定义
ConcurrentHashMap提供了5个构造器，这5个构造器最多也只是计算了下table的初始容量大小，并没有进行实际的创建table数组的工作：
ConcurrentHashMap采用了一种“懒加载”的模式，(这个在table属性定义时候的注释有描述)只有到首次插入键值对的时候，才会真正的去初始化table数组。
空构造器
/**
* Creates a new, empty map with the default initial table size (16).
*/
public ConcurrentHashMap() {
}
指定table初始容量的构造器
/**
* 指定table初始容量的构造器.
* tableSizeFor会返回大于入参（initialCapacity + (initialCapacity >>> 1) + 1）的最小2次幂值
* Creates a new, empty map with an initial table size
* accommodating the specified number of elements without the need
* to dynamically resize.
*
* @param initialCapacity The implementation performs internal
* sizing to accommodate this many elements.
* @throws IllegalArgumentException if the initial capacity of
* elements is negative
*/
public ConcurrentHashMap(int initialCapacity) {
if (initialCapacity < 0)
throw new IllegalArgumentException();
int cap = ((initialCapacity >= (MAXIMUM_CAPACITY >>> 1)) ?
MAXIMUM_CAPACITY :
tableSizeFor(initialCapacity + (initialCapacity >>> 1) + 1));
this.sizeCtl = cap;
}
根据已有的Map构造
/**
* 根据已有的Map构造ConcurrentHashMap.
* Creates a new map with the same mappings as the given map.
*
* @param m the map
*/
public ConcurrentHashMap(Map<? extends K, ? extends V> m) {
this.sizeCtl = DEFAULT_CAPACITY;
putAll(m);
}
指定table初始容量和负载因子的构造器
/**
* 指定table初始容量和负载因子的构造器.
* Creates a new, empty map with an initial table size based on
* the given number of elements ({@code initialCapacity}) and
* initial table density ({@code loadFactor}).
*
* @param initialCapacity the initial capacity. The implementation
* performs internal sizing to accommodate this many elements,
* given the specified load factor.
* @param loadFactor the load factor (table density) for
* establishing the initial table size
* @throws IllegalArgumentException if the initial capacity of
* elements is negative or the load factor is nonpositive
*
* @since 1.6
*/
public ConcurrentHashMap(int initialCapacity, float loadFactor) {
this(initialCapacity, loadFactor, 1);
}
指定table初始容量、负载因子、并发级别的构造器
concurrencyLevel只是为了兼容JDK1.8以前的版本，并不是实际的并发级别，loadFactor也不是实际的负载因子，这两个都失去了原有的意义，仅仅对初始容量有一定的控制作用
/**
* 指定table初始容量、负载因子、并发级别的构造器.
* 注意：concurrencyLevel只是为了兼容JDK1.8以前的版本，并不是实际的并发级别，loadFactor也不是实际的负载因子
* 这两个都失去了原有的意义，仅仅对初始容量有一定的控制作用
* Creates a new, empty map with an initial table size based on
* the given number of elements ({@code initialCapacity}), table
* density ({@code loadFactor}), and number of concurrently
* updating threads ({@code concurrencyLevel}).
*
* @param initialCapacity the initial capacity. The implementation
* performs internal sizing to accommodate this many elements,
* given the specified load factor.
* @param loadFactor the load factor (table density) for
* establishing the initial table size
* @param concurrencyLevel the estimated number of concurrently
* updating threads. The implementation may use this value as
* a sizing hint.
* @throws IllegalArgumentException if the initial capacity is
* negative or the load factor or concurrencyLevel are
* nonpositive
*/
public ConcurrentHashMap(int initialCapacity,
float loadFactor, int concurrencyLevel) {
if (!(loadFactor > 0.0f) || initialCapacity < 0 || concurrencyLevel <= 0)
throw new IllegalArgumentException();
if (initialCapacity < concurrencyLevel) // Use at least as many bins
initialCapacity = concurrencyLevel; // as estimated threads
long size = (long)(1.0 + (long)initialCapacity / loadFactor);
int cap = (size >= (long)MAXIMUM_CAPACITY) ?
MAXIMUM_CAPACITY : tableSizeFor((int)size);
this.sizeCtl = cap;
}
2.4 ConcurrentHashMap常量和属性
常量
/* ---------------- Constants -------------- */
/**
* The largest possible table capacity. This value must be
* exactly 1<<30 to stay within Java array allocation and indexing
* bounds for power of two table sizes, and is further required
* because the top two bits of 32bit hash fields are used for
* control purposes.
*/
private static final int MAXIMUM_CAPACITY = 1 << 30; //最大容量2的30次方
/**
* The default initial table capacity. Must be a power of 2
* (i.e., at least 1) and at most MAXIMUM_CAPACITY.
*/
private static final int DEFAULT_CAPACITY = 16; //默认容量 1<<4
/**
* The largest possible (non-power of two) array size.
* Needed by toArray and related methods.
*/
static final int MAX_ARRAY_SIZE = Integer.MAX_VALUE - 8;
/**
* The default concurrency level for this table. Unused but
* defined for compatibility with previous versions of this class.
*/
private static final int DEFAULT_CONCURRENCY_LEVEL = 16;
/**
* 负载因子，为了兼容JDK1.8以前的版本而保留。
* JDK1.8中的ConcurrentHashMap的负载因子恒定为0.75
* The load factor for this table. Overrides of this value in
* constructors affect only the initial table capacity. The
* actual floating point value isn't normally used -- it is
* simpler to use expressions such as {@code n - (n >>> 2)} for
* the associated resizing threshold.
*/
private static final float LOAD_FACTOR = 0.75f; //负载因子
/**
* The bin count threshold for using a tree rather than list for a
* bin. Bins are converted to trees when adding an element to a
* bin with at least this many nodes. The value must be greater
* than 2, and should be at least 8 to mesh with assumptions in
* tree removal about conversion back to plain bins upon
* shrinkage.
*/
static final int TREEIFY_THRESHOLD = 8; // 链表转树的阈值，即链接节点数大于8时， 链表转换为树.
/**
* The bin count threshold for untreeifying a (split) bin during a
* resize operation. Should be less than TREEIFY_THRESHOLD, and at
* most 6 to mesh with shrinkage detection under removal.
*/
static final int UNTREEIFY_THRESHOLD = 6; // 树转链表的阈值，即树节点树小于6时，树转换为链表.
/**
* 在链表转变成树之前，还会有一次判断：
* 即只有键值对数量大于MIN_TREEIFY_CAPACITY，才会发生转换。
* 这是为了避免在Table建立初期，多个键值对恰好被放入了同一个链表中而导致不必要的转化。
* The smallest table capacity for which bins may be treeified.
* (Otherwise the table is resized if too many nodes in a bin.)
* The value should be at least 4 * TREEIFY_THRESHOLD to avoid
* conflicts between resizing and treeification thresholds.
*/
static final int MIN_TREEIFY_CAPACITY = 64;
/**
* 在树转变成链表之前，还会有一次判断：
* 即只有键值对数量小于MIN_TRANSFER_STRIDE，才会发生转换.
* Minimum number of rebinnings per transfer step. Ranges are
* subdivided to allow multiple resizer threads. This value
* serves as a lower bound to avoid resizers encountering
* excessive memory contention. The value should be at least
* DEFAULT_CAPACITY.
*/
private static final int MIN_TRANSFER_STRIDE = 16;
/**
* 用于在扩容时生成唯一的随机数.
* The number of bits used for generation stamp in sizeCtl.
* Must be at least 6 for 32bit arrays.
*/
private static int RESIZE_STAMP_BITS = 16;
/**
* 可同时进行扩容操作的最大线程数.
* The maximum number of threads that can help resize.
* Must fit in 32 - RESIZE_STAMP_BITS bits.
*/
private static final int MAX_RESIZERS = (1 << (32 - RESIZE_STAMP_BITS)) - 1;
/**
* The bit shift for recording size stamp in sizeCtl.
*/
private static final int RESIZE_STAMP_SHIFT = 32 - RESIZE_STAMP_BITS;
/*
* Encodings for Node hash fields. See above for explanation.
*/
static final int MOVED = -1; // hash for forwarding nodes 标识ForwardingNode节点（在扩容时才会出现，不存储实际数据）
static final int TREEBIN = -2; // hash for roots of trees 标识红黑树的根节点
static final int RESERVED = -3; // hash for transient reservations 标识ReservationNode节点（）
static final int HASH_BITS = 0x7fffffff; // usable bits of normal node hash
/** Number of CPUS, to place bounds on some sizings */
static final int NCPU = Runtime.getRuntime().availableProcessors(); // CPU核心数，扩容时使用
属性
/* ---------------- Fields -------------- */
/**
* Node数组，标识整个Map，首次插入元素时创建，大小总是2的幂次.
* The array of bins. Lazily initialized upon first insertion.
* Size is always a power of two. Accessed directly by iterators.
*/
transient volatile Node<K,V>[] table;
/**
* 扩容后的新Node数组，只有在扩容时才非空.
* The next table to use; non-null only while resizing.
*/
private transient volatile Node<K,V>[] nextTable;
/**
* 计数基值,当没有线程竞争时，计数将加到该变量上。类似于LongAdder的base变量
* Base counter value, used mainly when there is no contention,
* but also as a fallback during table initialization
* races. Updated via CAS.
*/
private transient volatile long baseCount;
/**
* 控制table的初始化和扩容.
* 0 : 初始默认值
* -1 : 有线程正在进行table的初始化
* >0 : table初始化时使用的容量，或初始化/扩容完成后的threshold
* -(1 + nThreads) : 记录正在执行扩容任务的线程数
* Table initialization and resizing control. When negative, the
* table is being initialized or resized: -1 for initialization,
* else -(1 + the number of active resizing threads). Otherwise,
* when table is null, holds the initial table size to use upon
* creation, or 0 for default. After initialization, holds the
* next element count value upon which to resize the table.
*/
private transient volatile int sizeCtl;
/**
* 扩容时需要用到的一个下标变量.
* The next table index (plus one) to split while resizing.
*/
private transient volatile int transferIndex;
/**
* 自旋标识位，用于CounterCell[]扩容时使用。类似于LongAdder的cellsBusy变量
* Spinlock (locked via CAS) used when resizing and/or creating CounterCells.
*/
private transient volatile int cellsBusy;
/**
* 计数数组，出现并发冲突时使用。类似于LongAdder的cells数组
* Table of counter cells. When non-null, size is a power of 2.
*/
private transient volatile CounterCell[] counterCells;
// views 视图相关字段
private transient KeySetView<K,V> keySet;
private transient ValuesView<K,V> values;
private transient EntrySetView<K,V> entrySet;
网友评论